Python and backend automation
Deterministic jobs, command-line tools or services, retries, idempotency, structured logs, and scheduled execution.
Sweden-based independent operator · Remote B2B
I implement bounded Python automation, API integrations, PostgreSQL work, batch data pipelines, and regression harnesses. We first define the inputs, outputs, failure behavior, and acceptance test. Work begins only after a written scope and funded first milestone.
I am a Sweden-based independent operator publishing under Agentsor. The public repository is owned engineering proof—not client work, a customer case study, or evidence of prior customer results.
BoundedOne written first milestone
VerifiableObjective acceptance before coding
OperableTests, logs, setup, and handoff
Implementation scope
I quote only after the authorized interfaces, primary input and output, implementation artifacts, acceptance test, dependencies, and exclusions are clear.
Deterministic jobs, command-line tools or services, retries, idempotency, structured logs, and scheduled execution.
Authorized REST integrations, input validation, migrations, transaction-safe changes, and failure handling.
Parquet ingestion, schema checks, deterministic transforms, quarantine, and one agreed target adapter.
Fixture-driven pytest harnesses, pass/fail gates, machine-readable reports, CI integration, and handoff documentation.
Public owned proof
The public Parquet Guard repository demonstrates corrupt and schema-drift quarantine without stopping good files, contract-sensitive idempotency, damaged-output repair, cross-filesystem quarantine, structured logs, and per-file failure isolation.
Open the public GitHub repositoryThis is owned proof, not client work. It contains no customer data and proves only the behavior exercised by its 16 public tests and hosted CI workflow.
Commercial process
Describe the expected input, output, current stack, and pass/fail result without secrets, private links, or customer data.
I will say whether the boundary is serviceable and ask only the questions needed to quote it.
Scope, price, delivery date, dependencies, acceptance, and exclusions are written before implementation.
Delivery includes the agreed artifacts, tests, setup instructions, and operating notes.
Likely fit
A Python, API, PostgreSQL, data-pipeline, or regression-testing job with objective acceptance criteria, authorized interfaces, safe fixtures, and a fundable fixed first milestone.
Not a fit
Unpaid trials; staff augmentation; unauthorized production access; live evasion or security probing; anti-bot bypass; spam; prohibited scraping; payment-card handling; 24/7 staffed response; or guarantees of revenue, uptime, security, compliance, model validity, or business outcomes.
Recurring work only when real
After successful delivery, a separate monthly scope can be quoted only if a genuine repeating job exists, such as scheduled-run review, schema-change handling, or a bounded change allowance.
Any monthly scope names the cadence, response window, included time, exclusions, and termination boundary. It is never assumed from the first milestone.
Request one fit response
Email the expected input, output, current stack, objective pass/fail result, delivery window, and funded first-milestone budget band.
Do not include credentials, API keys, source code, client or customer names, personal data, production records, private links, or confidential material.
B2B inquiries only. Emailing requests one relevant response; it does not permit unrelated marketing or automatic follow-ups and creates no contract, reservation, charge, or obligation.